I would start with a vision, a set of principles, and a written description of what you’d like to accomplish. Then take that and work backwards based on the principles. For example if you hvae a principle of least surprise selecting something like GitLab might not make sense 😉